diff options
Diffstat (limited to 'benchmarks/dc14.dat')
| -rw-r--r-- | benchmarks/dc14.dat | 196 |
1 files changed, 196 insertions, 0 deletions
diff --git a/benchmarks/dc14.dat b/benchmarks/dc14.dat new file mode 100644 index 0000000..4719d9c --- /dev/null +++ b/benchmarks/dc14.dat @@ -0,0 +1,196 @@ +BEGIN NEW DATA CASE
+C BENCHMARK DC-14
+C 1st of 4 subcases will calculate constant-parameter equivalent network
+C AGNOSTIC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 { Ov 26
+C Add re-dimensioning request 1 April 2007 in order to regain those sparsity
+C diagrams showing non-zero structure. Until this past year, these were
+C missing for 3 * default dimensioning of Salford EMTP and also other versions
+C I suppose. Then, working to support Orlando Hevia and GNU F95 testing, a
+C change to NETEQV was made in January. Storage in KSPARS was changed from
+C INTEGER (1 character per 4/byte word) to CHARACTER*1. But, as Tsu-huei
+C discovered only toward the end of March (apparently testing was delayed),
+C Watcom died at the first attempted storage in KSPARS(1,1). Suspecting that
+C Watcom did not like the INTEGER vs. CHARACTER*1 mismatch of the final
+C argument, WSM retains the NETEQV code and simply replaces KARRAY within
+C SUBR26 by List 15 CSTO of MODELS. But, using 3 * default dimensioning,
+C LSPARS = 225 and this requires 225 ** 2 = 50625 cells which is more than
+C 3 * default will provide. So, as first tested, the two sparsity diagrams
+C were correctly missing from the output. But they were nice, and there is no
+C need to lose them. The dimensioned size of List 15 is substantially bigger,
+C so why not simply expand List 15 to 51K at the same time much more storage
+C is saved by reducing other list sizes (possible since they are not used,
+C anyway, for supporting programs only). "LABCOM tables = 234090 INTEGER
+C words" of the heading will shrink from 234090 to 84270. This is a novel
+C use of NLS for non-simulation, so has educational value in its own right.
+NEW LIST SIZES
+ 0 0 0 0 0 0 0 0 0 0
+ 0 0 0 0 51000 0 0 0 0 0
+ 0 0 0 { Note only List 15 = 51K differs from default size
+ 240000
+NETWORK EQUIVALENT { Request transfer to supporting program "NETEQV"
+C NITMAX IPRSUP EPSILN PERCEN IPNCOM NAME ---- Begin next sub-sub-case
+ 10 0 1.E-9 0.5 0 ABC?
+C K M X-OHMS
+ 1 2 0.5 { Branch node K, node M, reactance of branch
+BLANK card terminates branch cards (all branches to be excluded from equivalent)
+C BUS1 BUS2 X/R X-THEV NEGATE
+ 1DUMMY DUMMY 10. 0.1 0
+ 2DUMMY DUMMY 20. .05 0
+C $PUNCH { Flush punched cards pertaining to just-completed 2-bus case
+BLANK card ending data cases within "NETWORK EQUIVALENT". Exit to "OVER1"
+BEGIN NEW DATA CASE
+C 2nd of 4 subcases will calculate constant-parameter equivalent network
+NETWORK EQUIVALENT { Request transfer to supporting program "NETEQV"
+C NITMAX IPRSUP EPSILN PERCEN IPNCOM NAME ---- Begin next sub-sub-case
+ 10 0 1.E-9 1.0
+C K M X-OHMS
+ 1 2 278. { Branch node K, node M, reactance of branch
+ 1 3 87. { Branch node K, node M, reactance of branch
+ 1 4 145. { Branch node K, node M, reactance of branch
+ 1 5 145. { Branch node K, node M, reactance of branch
+ 1 6 5.16 { Branch node K, node M, reactance of branch
+ 1 7 467. { Branch node K, node M, reactance of branch
+ 1 8 316. { Branch node K, node M, reactance of branch
+BLANK card terminates branch cards (all branches to be excluded from equivalent)
+C BUS1 BUS2 X/R X-THEV NEGATE
+ 1DUMMY DUMMY 50. .09
+ 2DUMMY DUMMY 33. .15
+ 3DUMMY DUMMY 50. .09
+ 4DUMMY DUMMY 33. .15
+ 5DUMMY DUMMY 50. .09
+ 6DUMMY DUMMY 33. .15
+ 7DUMMY DUMMY 50. .09
+ 8DUMMY DUMMY 33. .15
+C $PUNCH { Flush the punched-card output of just-created equivalent network
+BLANK card ending data cases within "NETWORK EQUIVALENT". Exit to "OVER1"
+BEGIN NEW DATA CASE
+C 3rd of 4 data subcases was appended to illustrate realistic usage at BPA.
+C Data for BPA's EMTP was converted for solution using ATP by Eugene Davis of
+C of BPA during August of 1992. ATP advice from WSM. See Oct, 92, newsletter
+C ** HOT SPRINGS - TAFT - DWORSHAK EQUIVALENTS
+NETWORK EQUIVALENT, 9, { Max number of generators is reduced to yield diagram
+C NITMAX IPRSUP EPSILN PERCEN IPNCOM NAME --- 1st of 2 halves: zero seq
+ 10 0 1.E-9 1.0 0 ABC?
+C K M X-OHMS
+ 1 2 113.2 (K, M, REACTANCE OF BRANCH)
+ 2 3 194.0 (K, M, REACTANCE OF BRANCH)
+BLANK card ends branches that are to be excluded from the equivalent
+C BUS1 BUS2 X/R X-THEV NEGATE ---- Short circuit impedance at bus
+ 1DUMMY HOTS 9. 70.84 0 (zero-sequence values)
+ 2DUMMY TAFT 6. 48.51 0
+ 3DUMMY DWOR 9. 39.41 0
+C -- Done with Newton iteration # 1. Largest correction D4 = 3.84916323E-01
+C -- Done with Newton iteration # 1. Largest correction D4 = 1.33373074E-01
+C -- Done with Newton iteration # 1. Largest correction D4 = 5.54850941E-03
+C -- Done with Newton iteration # 1. Largest correction D4 = 1.96795976E-05
+C -- Done with Newton iteration # 1. Largest correction D4 = 1.14179987E-10
+C reactance computed after removal of user-flagged generators. .... LOOP = 2.
+C Gen X-new R-new X-fault X-thev X/R-thev
+C 1 1.19249599709E+02 1.32499555232E+01 7.08400E+01 7.08400E+01 9.000000
+C 2 8.22611517150E+01 1.37101919525E+01 4.85100E+01 4.85100E+01 6.000000
+C 3 4.66222308277E+01 5.18024786975E+00 3.94100E+01 3.94100E+01 9.000000
+C NITMAX IPRSUP EPSILN PERCEN IPNCOM NAME ----- End 0-sequence, begin +
+ 10 0 1.E-9 1.0 0 ABC?
+C K M X-OHMS K = "from" node number; M = "to" node number;
+ 1 2 30.33 X is reactance of branch not included in equiv.
+ 2 3 55.73 (branches represented explicitely in EMTP data)
+BLANK card ends branches that are to be excluded from the equivalent
+C BUS1 BUS2 X/R X-THEV NEGATE ---- Short circuit impedance at bus
+ 1DUMMY HOTS 16. 52.37 0 (positive-sequence values)
+ 2DUMMY TAFT 17. 38.59 0
+ 3DUMMY DWOR 19. 30.59 0
+C -- Done with Newton iteration # 1. Largest correction D4 = 6.33252721E-01
+C -- Done with Newton iteration # 2. Largest correction D4 = 3.90146597E-01
+C -- Done with Newton iteration # 3. Largest correction D4 = 1.39754125E-01
+C -- Done with Newton iteration # 4. Largest correction D4 = 5.79939902E-03
+C -- Done with Newton iteration # 5. Largest correction D4 = 2.06058080E-06
+C -- Done with Newton iteration # 6. Largest correction D4 = 1.96349525E-12
+C Final generator equivalent impedances, plus confirmation of original generator
+C parameters. Resistance is ignored during the internal computation, and is
+C added as a fixed percentage only at the end. X-fault is the short circuit
+C reactance computed after removal of user-flagged generators. .... LOOP = 1.
+C Gen X-new R-new X-fault X-thev X/R-thev
+C 1 1.55130207469E+02 9.69563796684E+00 5.23700E+01 5.23700E+01 16.000000
+C 2 9.81300314308E+01 5.77235479005E+00 3.85900E+01 3.85900E+01 17.000000
+C 3 4.10670198789E+01 2.16142209889E+00 3.05900E+01 3.05900E+01 19.000000
+C Note about order. Preceding data has two halves, with the zero sequence
+C sequence preceding the positive. This order is needed only for punched
+C cards that follow to be correct (zero sequence first). Otherwise, the
+C order is immaterial. In fact, the same solution is performed twice. It
+C is only data values that distinguish the two solutions.
+$PUNCH, dc14b.pch !
+C <++++++> Cards punched by support routine on 24-Aug-92 09.27.44 <++++++>
+C 51,DUMMYA,HOTS A, ,,, 1.324995552323771E+01, 1.192495997091393E+02,,,,,,
+C 52,DUMMYB,HOTS B, ,,, 9.695637966841947E+00, 1.551302074694711E+02,,,,,,
+C 53,DUMMYC,HOTS C, ,,,,,,,,,,
+C 51,DUMMYA,TAFT A, ,,, 1.371019195249209E+01, 8.226115171495253E+01,,,,,,
+C 52,DUMMYB,TAFT B, ,,, 5.772354790047830E+00, 9.813003143081311E+01,,,,,,
+C 53,DUMMYC,TAFT C, ,,,,,,,,,,
+C 51,DUMMYA,DWOR A, ,,, 5.180247869745662E+00, 4.662223082771096E+01,,,,,,
+C 52,DUMMYB,DWOR B, ,,, 2.161422098891424E+00, 4.106701987893705E+01,,,,,,
+C 53,DUMMYC,DWOR C, ,,,,,,,,,,
+BLANK card ending data subcases within NETWORK EQUIVALENT.
+BEGIN NEW DATA CASE
+C 4th of 4 subcases has nothing to do with the preceding 3. Instead,
+C it converts old Type-91, 92, or 93 switched elements to newer
+C pseudo-nonlinear branches.
+CHANGE SWITCH
+C Cards preceding the blank card ending branch cards are all discarded
+C except for the Type-91 time-dependent resistance. So, following a
+C dummy series R-L-C branch, let's insert one of these Type-91 branches:
+ NODE1 NODE2 1.0 { Series R-L-C will be discarded, of course
+91LEFT RIGHT 0.3E6 { From p. 20 example of 1980 Rule Book } 1
+ 0.0 300.
+ 0.3 200.
+ 0.6 150.
+ 1000. 150.
+ 9999
+91 NODE1 NODE2 COPYL COPYR .25E6 { Final card from p. 20 of 1980 Rule Book
+BLANK card ending program branch cards. This is flag that switches follow.
+C Type-92 switched-R element follows (to be converted):
+C 3456789012345678901234567890123456789012345678901234567890
+92JDAYA LMONA 8.5 3.E5 3.5E5 3
+C Type-93 switched-L element follows (to be converted):
+93SENDA NEUTRL 4.2 2.5 0.7 3.3 1
+BLANK card terminating program switch cards
+C Since all cards between the blank card ending switch cards (immediately
+C above) and the start of a new case ("BNDC" in unabbreviated, long form)
+C are to be discarded, we will not show anything here. No intelligence
+C is involved in the "CHANGE SWITCH" logic, so nothing can be tested.
+BEGIN NEW DATA CASE
+$PUNCH { Flush the punched-card output of equivalent pseudo-nonlinear elements
+C 91LEFT RIGHT 3333.
+C 0.3E6
+C 0.0 300.
+C 0.3 200.
+C 0.6 150.
+C 1000. 150.
+C 9999
+C 91 NODE1 NODE2 COPYL COPYR 3333.
+C 99JDAYA LMONA 0 3 2 035
+C 0.0000001 3.E5
+C 1.0 0.300008500E+06
+C 9999
+C 98SENDA NEUTRL 1.0 3.3
+C 0.1320000E+04 3.3
+C 0.2640000E+04 0.4224000E+01
+C 9999
+C Because this was not the first subcase of the disk file that punched
+C cards ($PUNCH usage), note that the usual $ERASE was not required
+C at the start. The $PUNCH at the end of the preceding subcase left
+C the punch buffer empty. But the final $PUNCH (immediately before
+C these comment cards) requires special attention. Were there not at
+C least five data cards remaining at the end of the final meaningful
+C data case, the blank terminating case would not even be read (the
+C program would instead discard remaining cards and jump out to issue
+C the initial prompt once again. The present set of comment cards do
+C number more than 5, so they should protect against this possibility
+C provided comment cards were not being destroyed by NOCOMM = 1 as
+C set in STARTUP. But we can not be sure of this, so it is safer to
+C have 5 or more non-blank data cards. By adding 2 unused blank cards
+C below, BNDC + $PUNCH + original blank + these 2 new ones total 5.
+C The blank bounding case, including the imbedded $PUNCH (which we
+C need to flush the final punched cards) will be read.
+BLANK
+BLANK { 1st of 2 unused lines can have any content. This is line 4 of 5.
+BLANK { 2nd of 2 unused lines can have any content This is line 5 of 5.
|